BRIEFING PAPER Number 8531, 22 October 2020 Preparing for the 2021 By Cassie Barton

census ( and )

Contents: 1. Background to the census 2. Legislative process 3. Content of the 2021 census 4. How the 2021 census will work 5. Data processing and outputs

www.parliament.uk/commons-library | intranet.parliament.uk/commons-library | [email protected] | @commonslibrary 2 Preparing for the 2021 census (England and Wales)

Contents

Summary 4 What will the census ask about? 4 An online census 4 Covid-19 and the census 5 Census legislation 5 Outputs and analysis 5 1. Background to the census 6 1.1 History of the census in the UK 6 1.2 Why do a census? 7 1.3 Who is responsible for the census? 7 1.4 Decision to go ahead with a 2021 census 7 1.5 Plans for the 2021 census 9 2. Legislative process 10 2.1 Primary and secondary legislation 10 The Census Act 1920 10 The Census (Return Particulars and Removal of Penalties) Act 2019 11 The Census Order 11 The Census Regulations 11 2.2 The census in and Northern Ireland 12 2.3 EU legislation 13 3. Content of the 2021 census 14 3.1 Choosing census topics 14 3.2 New questions 16 Gender identity 17 Sexual orientation 19 Past service in the UK Armed Forces 21 3.3 Repeating questions 22 Ethnic group 22 Marital and civil partnership status 26 National identity 26 Unpaid care 26 Long-term health and disability 27 Qualifications 27 Economic activity 27 3.4 Discontinued questions 27 Number of rooms 27 4. How the 2021 census will work 28 4.1 An online census 28 4.2 Population covered by the census 28 4.3 Covid-19 and the census 29 4.4 Operation of the census 30 Staffing 30 Identifying and contacting respondents 30 Coverage and non-response 31 Census rehearsal 32 4.5 Engagement and support 32 4.6 Costs 35 3 Commons Library Briefing, 22 October 2020

5. Data processing and outputs 36 5.1 Data processing, analysis and security 36 Processing census data 36 Data security 36 Linking to administrative sources 36 5.2 Dissemination 37 Appendix 1: History of census topics 38

Cover page image copyright ESA, CC BY-SA 3.0 IGO [cropped]

4 Preparing for the 2021 census (England and Wales)

Summary

A population census takes place in the UK every ten years. The next census in England and Wales will take place on 21st March 2021, and will be administered by the Office for National Statistics (ONS). The census seeks to collect demographic information from all households and communal establishments. The 2021 census will be the first to be carried out primarily online. Census data provides a level of detail that isn’t possible from other government surveys – information is available about small population groups, and for small geographic areas. Census data contributes to policy decisions, including local government funding allocations, and provides a benchmark for other official statistics. What will the census ask about? The census collects a wide range of demographic information about topics like age, sex, marital status, health, education and housing. There will be three new questions in the 2021 census, covering: • Veteran status: whether the respondent has ever served in the UK Armed Forces. • Sexual orientation: whether the respondent identifies as gay, lesbian, bisexual, heterosexual, or some other sexual orientation. This question will be voluntary and only asked of respondents aged 16 and over. • Gender identity: whether the respondent’s gender is different from the sex they were registered as at birth. This question will also be voluntary and limited to respondents aged 16 and over. Almost all of the topics asked about in 2011 will appear again in 2021, although the way in which some questions are asked will change. The ethnic group question will include a new tick-box for people of Roma ethnicity, alongside the existing ‘Gypsy or Irish Traveller’ tick-box. An online census The 2021 census will be the first time that the census is conducted primarily online. The ONS’ target is to have 75% of census returns completed online, with the remainder completed on traditional paper forms. Most households will receive a unique code in the post which will allow them to complete their census returns online. The ONS will put provisions in place to support households that may have difficulty accessing an online census. Households in some areas will receive a paper form to start with, and any household will be able to receive a paper form on request. The ONS also plans to provide in- person support sessions in some locations (e.g. in public libraries). 5 Commons Library Briefing, 22 October 2020

Covid-19 and the census The Covid-19 outbreak poses operational challenges for preparing and carrying out a census. The Scottish Government announced in July 2020 that it would delay Scotland’s census until March 2022 for this reason. A census will still take place in March 2021 in England and Wales, as it will in Northern Ireland. The ONS has published an operational planning response that sets out how it plans to address the impact of the pandemic on its census operations. For example, it plans to shift more of its community engagement activity online. Census legislation All of the necessary legislation for a census to take place in England and Wales has now come into force. The Census Act 1920 is the current statutory authority for a census to be taken in England and Wales. Secondary legislation must also be passed for a census to take place. A Census Order specifies the topics asked about, and Census Regulations set out operational details. These pieces of secondary legislation have now become law. In October 2019, the government passed the Census (Return Particulars and Removal of Penalties) Act 2019 which adds sexual orientation and gender identity to the list of topics which the census can ask about, and to ensure that there is no penalty for respondents who don’t answer these questions. Census legislation is devolved in Scotland and Northern Ireland, and separate censuses are administered by their respective governments. There is agreement between the statistical offices of the UK nations that census statistics should be harmonised where possible. Outputs and analysis The ONS has said that it will make better use of administrative data to enhance the findings of the 2021 census. It plans to use administrative data on property size in place of a question about the number of rooms that respondents have in their home. The ONS aims to publish an initial set of census reports one year after census day, and to make all other outputs available within two years. It plans to develop a flexible online dissemination system that will let users select the specific data they require.

Section 1 of this briefing gives background to the decision to carry out a census in 2021. Section 2 explains the legislative process behind the census. Sections 3 and 4 explain the proposed content and implementation of the census, respectively. Section 5 discusses the proposals for data processing and census outputs. 6 Preparing for the 2021 census (England and Wales)

1. Background to the census 1.1 History of the census in the UK ’s first attempt to measure its population through a census took place in 1801. The census was led by John Rickman, who had outlined the advantages of taking a census in an essay the year before. He argued that understanding the size of the population would allow the country to plan corn production and military conscription more effectively, and that a census would “generally encourage the social sciences to flourish”. The case for a census was taken up by several MPs, and a Census Bill passed quickly through Parliament. Rickman ended up overseeing every decennial census between 1801 and 1831.1 A census has taken place every ten years in Great Britain (i.e. England, Scotland and Wales) since 1801, with the exception of 1941, when the census was prevented by the Second World War. The census of 1841 is often described as the first modern census because it recorded key details including names and occupations, and was carried out systematically by enumerators. Decennial censuses were taken in Ireland between 1841 and 1911. The first Northern Ireland census was in 1926, followed by another in 1937, with censuses taking place every ten years from 1951. New topics have been added to the census over the years, while others have fallen away. Appendix 1 summarises how the content of the census has changed since 1841. It was possible to complete the England and Wales census online in 2011, but 2021 is the first year in which the census will be carried out primarily online.

Census history: key dates 2 1801 The first census of Great Britain is carried out. 1841 For the first time the census requires individual households to complete forms and return them to enumerators. 1911 Suffragettes stage a boycott of the census; activist Emily Wilding Davison is found hiding in the Houses of Parliament on census night. 1920 The Census Act passes, allowing for a census to be taken every ten years without further primary legislation. 1941 The decennial census is missed for the first time due to the Second World War. 1961 A computer is used to process census results for the first time. 1966 An additional census is taken, requiring information from only a sample of the population (approximately 10%). 1991 The census asks about ethnic group for the first time. 2001 New questions are asked about religion and general health. 2011 Census returns can be completed online for the first time. New topics include passports held, national identity and language.

1 Office for National Statistics (1993), The 1991 Census: A User’s Guide 2 Ibid. and Office for National Statistics (1977), Guide to Census Reports, Part 2. Accessed 18 March 2019 via http://www.visionofbritain.org.uk/census/Cen_Guide/2 7 Commons Library Briefing, 22 October 2020

1.2 Why do a census? Government, Parliament, local authorities, the health service, the education and academic communities, businesses and the public use the census to give them reliable information on the number and characteristics of people and households. The census is unique in the level of detail and consistency it provides, as it aims to cover every person usually resident in the UK on one night. Other surveys and sources are typically limited in that they cover only a sample of the population, and/or don’t cover all of the same topics as the census does simultaneously. Uses for census data include: • funding formulae used to allocate central government resources to other organisations such as the devolved administrations (via the Barnett formula), local authorities and health bodies. • policy development and planning by central, local and regional government, in areas including equalities, housing, transport, employment and health. • census data continues to underlie much social research, both within and outside government, and businesses use it to target their marketing and location policies. Government researchers use census data to assess the representativeness of other social surveys.

1.3 Who is responsible for the census? The Statistics Authority (UKSA), the arm’s-length public body responsible for official statistics in the UK, holds the statutory powers to conduct a census in England and Wales. The Office for National Statistics (ONS) acts as the executive arm of UKSA and carries out the planning and operation of the census. Section 2 explains the legislative framework behind the census in more detail.

1.4 Decision to go ahead with a 2021 census In the run-up to the 2011 census concerns were raised about its cost- effectiveness, and whether similar data could be more efficiently provided by surveys and government-held administrative data. The Treasury Select Committee report Counting the population in 2008 evaluated the suitability of the census for population statistics and recommended that: the Statistics Authority set strategic objectives to ensure that the data gathered throughout the UK can be used to produce annual population statistics that are of a quality that will enable the 2011 Census to be the last census in the UK where the population is counted through the collection of census forms.3 In 2010, UKSA commissioned a review of the future of population statistics from the ONS. This review included a public consultation

3 House of Commons Treasury Committee, Counting the population, HC 183–I 2007– 08, May 2008 8 Preparing for the 2021 census (England and Wales)

launched in 2013, which proposed two options for the census: continuing to take a decennial census but with forms filled out primarily online, or ‘using existing government data and compulsory annual surveys’ to generate equivalent statistics.4 The ONS published a response to the consultation in March 2014. It concluded that there was demand from a range of stakeholders for data on small geographic areas and populations, as well as concerns from these stakeholders that survey and administrative data would not meet their needs in this regard.5 In March 2014 the then National Statistician, Jil Matheson, published her recommendation for the future of the census: • An online census of all households and communal establishments in England and Wales in 2021 as a modern successor to the traditional, paper-based decennial census. ONS recognises that special care would need to be taken to support those who are unable to complete the census online. • Increased use of administrative data and surveys in order to enhance the statistics from the 2021 Census and improve annual statistics between censuses. 6 A similar conclusion was reached by the Public Administration Select Committee in its report Too soon to scrap the Census, published in April 2014. It concluded that “the alternative options for the collection of population statistics are not sufficiently advanced to provide a proper replacement […] however, the increasing cost and deficiencies of a traditional census must be recognised”. 7 The government endorsed the National Statistician’s proposed approach in a letter from the then Cabinet Office Minister to the Chair of UKSA in July 2014: We agree with the recommendation for an online census in 2021 as a modern successor to the traditional paper-based decennial census, with support for those who are unable to complete the census online. We welcome the increased use of administrative data in producing the census in 2021 and other population statistics, and to improve statistics between censuses, since this would make the best use of all available data and provide a sound basis for the greater use of administrative data and surveys in the future. 8 The letter also noted that the government continued to hold ambitions for an administrative-data census after 2021: However, our support for the dual running of an online (decennial) census with increased use of administrative data is

4 ONS, The census and future provision of population statistics in England and Wales: report on the public consultation, March 2014 5 Ibid. 6 ONS, The census and future provision of population statistics in England and Wales: recommendation from the National Statistician and Chief Executive of the UK Statistics Authority, March 2014 7 House of Commons Public Administration Select Committee, Too soon to scrap the census, HC1090 2013-14, April 2014 8 Letter, Rt Hon Sir Francis Maude MP to Sir Andrew Dilnot, 18 July 2014 9 Commons Library Briefing, 22 October 2020

only relevant to 2021 and not for future censuses. Our ambition is that censuses after 2021 will be conducted using other sources of data and providing more timely statistical information. 9

1.5 Plans for the 2021 census The government published a White Paper in December 2018, titled Help shape our future: the 2021 Census of population and housing in England and Wales. The White Paper sets out the ONS’ recommendations for the content and scope of the census in England and Wales, the way in which it will be implemented, and the proposed timescale for secondary legislation. The necessary legislation to conduct a census in 2021 is now in force (see section 2 of this briefing). The ONS has confirmed that the 2021 census will be conducted primarily online. The ONS plans to support the 2021 census with government administrative topics on certain topics (number of rooms and income). The White Paper also refers to a programme of work by ONS to base its population and migration statistics primarily on administrative data: ONS will consult with users on how these administrative data can come to the forefront of the statistical system and ensure that it is ready to make recommendations to government in 2023 on the future of the census. 10 This briefing paper explains more about the proposed content, operation and output of the census in sections 3, 4 and 5 respectively.

9 Ibid. 10 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018) p. 118 10 Preparing for the 2021 census (England and Wales)

2. Legislative process 2.1 Primary and secondary legislation This section sets out the legislation underpinning the census in England and Wales. For information on the census in Scotland and Northern Ireland, see the next section. The primary legislation for the census in England and Wales is the Census Act 1920. Each census requires two pieces of secondary legislation in order to happen: a Census Order and Census Regulations. The Census Order sets out the topics to be covered (as well as other details including the date), while the Census Regulations set out the operational procedure for taking the census and include a sample questionnaire. The National Assembly for Wales (Transfer of Functions) (No. 2) Order 2006 makes provision for the Welsh Parliament to be consulted about the content of the Census Order, and gives it the power to make Census Regulations for Wales. 11 The Census Act 1920 The current statutory authority for taking a census of population in England and Wales is the Census Act 1920.12 Section 1 of the Census Act 1920 enables an Order to direct the taking of a census for Great Britain or any part of it. In England and Wales, the duty to carry out the census in section 2 of the 1920 Act (formerly conferred on the Registrar General) was transferred to the Statistics Board by the Statistics and Registration Service Act 2007 (c. 18, Schedule 1, paragraphs 1 and 2). 13 The United Kingdom Statistics Authority (UKSA) holds the statutory powers of the Statistics Board and acts through the Office of National Statistics as its executive office (established under section 32 of the 2007 Act) in carrying out the work of preparing for, and undertaking a census. The Schedule to the Census Act 1920 sets out the topics that the census can ask about in England and Wales: (1) Names, sex, age (2) Occupation, profession, trade or employment (3) Nationality, birthplace, race, language (4) Place of abode and character of dwelling (5) Condition as to marriage or civil partnership, relation to head of family, issue born in marriage (5a) Religion (5b) Sexual orientation. (5c) Gender identity.

11 The National Assembly for Wales (Transfer of Functions) (No. 2) Order 2006, sections 3 and 4 12 Census Act 1920 (10 & 11 Geo. 5, ch. 41) 13 Statistics and Registration Service Act 2007 (ch. 18) 11 Commons Library Briefing, 22 October 2020

(6) Any other matters with respect to which it is desirable to obtain statistical information with a view to ascertaining the social or civil condition of the population14, 15 The Census (Return Particulars and Removal of Penalties) Act 2019 The ONS recommended introducing new voluntary questions on sexual orientation and gender identity in its 2018 White Paper. To enable this, the Census (Return Particulars and Removal of Penalties) Act 2019 was passed in October 2019. The 2019 Act amended the 1920 Act to add sexual orientation and gender identity to the Schedule of particulars that the census may ask about in England and Wales, and amended Section 8(1A) of the 1920 Act such that no-one is liable for a criminal penalty for not answering questions on this topic. Primary legislation is not necessary to introduce new topics to the census; topics can also be introduced under paragraph (6) of the existing Schedule (see above). Introducing new topics through primary legislation allows the government to clarify that the questions are voluntary and penalties do not apply to non-responders. 16 The Census (Return Particulars and Removal of Penalties) Bill [HL] 2017- 19 was introduced in the House of Lords on 1 May 2019 and passed its Lords stages without amendment on 26 June 2019. Its Commons stages also passed without amendment and it received Royal Assent on 8 October 2019. Section 3.2 of this briefing also provides more background on the sexual orientation and gender identity questions. The Library briefing for the Commons stages of the Bill (CBP 8605) explains the legislation and its reception in more detail. The Census Order Section 1 of the Census Act 1920 permits the laying of an Order in Council directing that a census be taken and detailing its contents. The Census Order is a piece of secondary legislation that sets out the date of the census, the topics the census will ask about, and the population to be covered by the census. The Census (England and Wales) Order 2020 was laid in draft before Parliament on 2 March 2020 and was made (became law) on 20 May 2020. The Census Regulations Section 3 of the Census Act 1920 permits the laying of Regulations that detail the procedures for undertaking the census and contain copies of

14 Census Act 1920 (10 & 11 Geo. 5, ch. 41), Schedule: Matters in respect of which particulars may be required. 15 The Census (Amendment) (Scotland) Act 2019 amended the Census Act 1920 differently as it applies in Scotland. In Scotland, item (5b) refers to transgender status and history and item (5c) refers to sexual orientation. 16 Census (Return Particulars and Removal of Penalties) Bill [HL] Explanatory Notes (1 May 2019) 12 Preparing for the 2021 census (England and Wales)

the relevant census forms. Regulations have been made separately in the UK Parliament (for England) and the Welsh Parliament. The Regulations cover operational details, including the appointment and duties of census staff, follow-up and non-compliance procedures, and security and confidentiality. The Regulations also cover the census questionnaire itself, and include the layout of the paper questionnaire and a description of the online questionnaire (including final wording for the questions). The Census (England) Regulations 2020 came into force on 23 June 2020 and the Census (Wales) Regulations 2020 came into force on 26 June 2020.

2.2 The census in Scotland and Northern Ireland The census is a devolved matter in Scotland and Northern Ireland. The census became a devolved matter in Scotland following the introduction of the Scotland Act 1998 and the Census Act 1920 now applies separately there. 17 In July 2020, the Scottish Government announced that Scotland’s census would be moved to March 2022 instead of March 2021 due to the impact of Covid-19 on census preparations (see section 4.3 of this briefing for more on this). 18 Information about plans for the 2022 census is available on the Scotland’s Census website. Separate legislation covers the census in Northern Ireland – the Census Act (Northern Ireland) 1969. 19 The Northern Ireland Statistics and Research Agency (NISRA) still plans to carry out a census in 2021, and has published details about its plans on the NISRA website. The UK government’s White Paper includes the following statement on harmonisation of the censuses between nations: UK harmonisation will be achieved through close liaison and co- operation between the three census offices. A statement of agreement has been drawn up between the National Statistician and the Registrars General for Scotland and Northern Ireland governing arrangements for the 2021 censuses in the UK. This statement ensures that the independent censuses carried out in each constituent country of the UK provide consistent and coherent statistics that are capable of meeting users’ requirements. 20 The statement of agreement was published in 2015 and is available to read online. 21

17 Scotland Act 1998 (ch. 46) 18 News release – Scotland’s Census to be moved to March 2022, 17 July 2020 [Accessed 7 October 2020] 19 Census Act (Northern Ireland) 1969 (ch. 8) (Act of the Northern Ireland Parliament) 20 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018) p. 115 21 The conduct of the 2021 Censuses in the UK: A statement of agreement between the National Statistician and the Registrars General for Scotland and Northern Ireland (October 2015) 13 Commons Library Briefing, 22 October 2020

2.3 EU legislation Regulation (EC) No 763/2008 of the European Parliament and Council requires member states to provide statistics about the population and housing based on a decennial census. The 2011 census, along with some existing survey information, helped the United Kingdom to meet its obligations under this Regulation in 2011. 22 The government has made a Statutory Instrument to revoke this regulation following the UK’s departure from the European Union. The regulation does not underpin the UK’s domestic legislation regarding the taking of a decennial census, so revoking the regulation does not directly affect the 2021 census.23

22 OJ L 218, 13 August 2008 23 Office for National Statistics and Cabinet Office, Explanatory Memorandum to the UK Statistics (Amendment Etc.) (EU Exit) Regulations 2019, 24 January 2019 14 Preparing for the 2021 census (England and Wales)

3. Content of the 2021 census

The content of the questions in the 2021 census has now been finalised. This section outlines how the ONS chooses topics to ask about, and explains new, amended and discarded questions in more detail. The questions and response options are set out in secondary legislation which has now come into force. The full questionnaire (both online and paper versions) can be read in Schedule 2 of the Census (England) Regulations 2020 and a sample paper form is available for download from the ONS website. The table overleaf summarises the topics that the 2021 census will cover, compared with the topics covered in 2011. Appendix 1 summarises the topics covered by the census since 1841. Not all topics are asked of all respondents. The census questions will be used to derive information about related topics, including short- and long-term international migration, patterns of travel to work, and National Statistics Socio-Economic Classification (NS-SEC).

3.1 Choosing census topics The ONS launched a consultation on the content of the 2021 census in June 2015, seeking input from stakeholders on the topics that the census should cover. A response to the consultation was published in May 2016, evaluating the suggested topics against a set of criteria.24 A set of topic-specific reports was published alongside the main response, each of which includes a summary of how user requirements were evaluated – these are available for download from the ONS website. The White Paper explains how the ONS selects topics to be included in the census: The topics proposed for the census are those that have a strong and clearly-defined user need and where questions have been devised that can be expected to produce reliable and accurate data. Moreover, in each case there is no other comparable and accessible source of the information able to meet the need. 25 The ONS uses a set of detailed criteria to assess user need, also explained in the White Paper. 26 These criteria include a need for data on small geographical areas or populations, as well as a need for analysis that combines multiple topics. Topics that don’t have these requirements are considered to be better-served by surveys or administrative data. The purpose of the data is also considered, as is continuity with previous censuses. Additionally, the ONS considers

24 ONS, The 2021 census – assessment of initial user requirements on content for England and Wales: response to consultation, May 2016 25 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018) p. 26 26 Ibid, p. 29 15 Commons Library Briefing, 22 October 2020

public acceptability (i.e. whether the topic would be considered too intrusive to ask about) and respondent burden (i.e. whether the census overall would be too long or difficult to reasonably complete). The ONS’ operational requirements, including the cost of coding responses and space on the paper form, are also considered. Proposed new questions are developed and tested to find out whether they will provide accurate and reliable results in census conditions. The ONS published summaries of this research in 2017 and 2018. In June 2020, the ONS published an in-depth explanation of its question and questionnaire development process. The ONS has also published a series of reports on question development for specific topics. 16 Preparing for the 2021 census (England and Wales)

Topics covered by the 2021 census in England and Wales

Residents in households and communal establishments Households only Topic 2011 2021 Topic 2011 2021 Name   Household / family relationships   Sex   Accommodation type   Date of birth   Self-contained accommodation   Marital or civil partnership status   Number of rooms   Ethnicity   Number of bedrooms   National identity   Central heating type   Sexual orientation*†   Tenure / landlord type   Gender identity*†   Number of cars / vans   Veteran status   Amount of unpaid care provided   Communal establishments only Health status   Type of establishment   Long-term illness or disability   Position within the establishment   Qualifications   Country of birth   Notes Month/year of entry into UK   * Voluntary questions Intended length of stay in UK   † 16+ year olds only Address one year ago   ‡ Wales only Passports held   Religion*   Welsh language skills‡   Main language   English-language proficiency   Economic activity   Occupation   Industry / name of employer   Transport to place of work   Supervisor status   Workplace address   Hours worked   Second residence   Students' term-time address  

3.2 New questions There will be three new questions on the 2021 census, on past service in the UK Armed Forces, gender identity, and sexual orientation. The latter two questions will be voluntary. In order to ensure that the questions could be asked on a voluntary basis, the government passed the Census (Return Particulars and Removal of Penalties) Act 2019 (see section 2.1 of this briefing for more). The UK Armed Forces question will not be voluntary, and as such does not require primary legislation to be introduced to the census. 17 Commons Library Briefing, 22 October 2020

Gender identity The 2021 census will include a question about gender identity, asking respondents whether their gender is the same as the sex they were registered as at birth. This question will only be asked of people aged 16 or over, and will be voluntary. The question will be separate from the question about sex (i.e., whether the respondent is male or female), which will be phrased in the same way as previous years. This question is designed to collect data about the size of the transgender population, as there is currently no official data on this (the word ‘transgender’ is used here to describe people whose gender identity does not match the sex they were registered as at birth). The Government Equalities Office (GEO) has said that there may be 200,000 to 500,000 transgender people in the UK, but stresses that we don’t know the true population because of the lack of robust data.27 The ONS has identified user need for official estimates in order to support policy- making and monitor equality duties. The ONS has also identified a need for census respondents to be able to provide accurate information about their gender identity, and considers that the existing sex question doesn’t meet this need for everyone. The ONS has carried out research evaluating different approaches to asking about gender identity (described in its 2017 and 2018 research updates). The ONS published a full question development report in June 2020. As part of its research, the ONS considered asking a single question on sex with a third option for those who do not identify as male or female, but concluded that doing so would risk the quality of existing data on sex. The ONS also found that a modified sex question would not be acceptable to some groups. In a letter to the Women and Equalities Select Committee, Iain Bell, the Deputy National Statistician for Population and Public Policy explained: Testing found that, overall, the addition of a third option meant individuals were unclear how to answer and the quality of male and female data was reduced. Our stakeholder engagement activity also uncovered some very strong views around the acceptability of adding a third response option, and that the sex question should remain as a binary option. 28 The letter followed an oral evidence session held by the committee in February 2019, at which Iain Bell commented: Our work had to come through and navigate a path that enabled the LGBT community to identify and feel included in the census, without having other groups feel as though they could not take part in the census because of it. 29 A large-scale survey found that most respondents find a question on gender identity acceptable (80% in England and 75% in Wales would

27 Government Equalities Office, Trans people in the UK, 2018 28 Letter from Iain Bell, Deputy National Statistician, to Maria Miller MP, Chair, Women and Equalities Select Committee, 12 March 2019 29 Women and Equalities Committee, Oral Evidence: Preparations for the 2021 Census, HC 1909, 6 February 2019 18 Preparing for the 2021 census (England and Wales)

find the question acceptable; 99% would complete the census if it included this question). 30 The ONS also found that “adding a caveat to the sex question that a gender identity question will follow the sex question makes the question more acceptable to non-binary and transgender people”.31 How will the census ask about sex and gender identity? The wording of the sex and gender identity questions is set out in the Census (England) Regulations (2020) and the equivalent regulations in Wales. Both sets of regulations have now come into force, which means that the questions will be phrased as follows. Note that the gender identity question will not directly follow the sex question, but instead appears later in the questionnaire. What is your sex? A question about gender identity will follow if you are aged 16 or over. ⬜ Female ⬜ Male

Is the gender you identify with the same as your sex registered at birth? This question is voluntary. ⬜ Yes ⬜ No, write in gender identity: [ ]32 Concerns have been raised about how transgender people, non-binary people (those who don’t identify as having a male or female gender) and intersex people (those born with biological sex characteristics that don’t meet the typical definition of male or female) should answer the sex question on the census. Baroness Barker raised this during the House of Lords stages while the 2019 Act was progressing through Parliament: [T]rans people have for many years been filling in the census and have done so in their lived-in identity. Is it anticipated that that will happen from now on and that a trans person will respond in their lived-in identity? […] Secondly, what does the Government expect non-binary people to do? […] [Intersex people] are currently assigned a sex at birth to have their birth recorded. Subsequently their sex may be changed. What are those people supposed to do?33 Lord Young of Cookham, then the Lords Spokesperson for the Cabinet Office, responded to these concerns in a letter to Baroness Barker: The guidance ONS is currently developing for the binary sex question is intended to be more comprehensive than that provided for 2011 Census, in order to assist those who identify as non-binary, trans, and/or with variations in sex characteristics, to

30 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018), pp. 37-38 31 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018), p. 38 32 The Census (England) Regulations 2020 (SI 2020/560), Schedule 2, part 4; The Census (Wales) Regulations 2020, (SI 2020/55 W. 128), Schedule 2, part 4. 33 HL Deb 13 May 2019 c1433 19 Commons Library Briefing, 22 October 2020

answer the questions. While the 2021 Census guidance will be more expansive, and reflect the inclusion of the new questions, the current underlying aim is to collect comparable data to 2011 and earlier censuses. In 2011, respondents who were unsure how to answer the sex question were advised that they need not answer it according to the sex on their birth certificate, whether or not they had a Gender Recognition Certificate. The current draft guidance is consistent with this. 34 The ONS has since published a draft of the guidance to accompany the sex and gender identity questions, which was used in the autumn 2019 census rehearsal (see section 4.3 for more on the rehearsal). The draft guidance to accompany the sex question is as follows: If you are one or more of non-binary, transgender, have variations of sex characteristics, sometimes also known as intersex, the answer you give can be different from what is on your birth certificate. If you’re not sure how to answer, use the sex registered on your official documents, such as passport or driving licence, or whichever answer best describes your sex. A later question gives the option to tell us if your gender is different from your sex registered at birth, and, if different, to record your gender. 35 This guidance is similar in content to guidance given in 2011. The 2011 census website included guidance for each question, with the following given for the sex question: Please select either male or female for your sex. Transgender or transsexual: select the answer which you identify yourself as. You can select either ‘male’ or ‘female’, whichever you believe is correct, irrespective of the details recorded on your birth certificate. You do not need to have a Gender Recognition Certificate. 36 Sexual orientation There will also be a new question about sexual orientation. As with the gender identity question, this question will be optional and will only be asked of respondents aged 16 or over. The question will be phrased as follows: Which of the following best describes your sexual orientation? This question is voluntary. ⬜ Straight/Heterosexual ⬜ Gay or Lesbian ⬜ Bisexual

34 Letter from Lord Young of Cookham to Baroness Barker, 21 May 2019 (DEP 2019- 0575) 35 ONS, Guidance for questions on sex, gender identity and sexual orientation for the 2019 Census Rehearsal for the 2021 Census, 11 September 2019 36 ONS, 2011 Census website, How to complete your census: question 2 [24 May 2011, Archived 2 August 2011] 20 Preparing for the 2021 census (England and Wales)

⬜ Other sexual orientation, write in: [ ]37 The ONS has identified a user need for better data on sexual orientation – particularly for small areas – to inform policy-making and service provision, as well as monitoring equality duties. The ONS has previously used the Annual Population Survey (APS) to estimate the size of the lesbian, gay and bisexual (LGB) population in the UK. According to these estimates, around 1.2 million people identified as LGB in 2018 (around 2% of the population). 38 However, the sample population used in the APS isn’t big enough to provide robust estimates of the LGB population in smaller areas. 39 The ONS tested the sexual orientation question in 2017. The research found that people who were given the sexual orientation question were not substantially less likely to complete the census. The distribution of responses was similar to the distribution found in the APS. A large-scale survey of households also found that most respondents found it acceptable to include this question (70% said it was acceptable; 80% said they would provide an accurate answer for themselves and on behalf of other household members). 40 In June 2020, the ONS published a report summarising its development of the sexual orientation question. The LGBT campaign group Stonewall issued a press release in 2016 when the proposals for sexual orientation and gender identity questions were first announced: Accurate population data on sexual orientation and gender identity allows organisations to develop services and initiatives which are targeted to the needs of their LGBT employees and/or service-users and local communities, and in doing so meet their requirements under the Equality Act 2010 […] But before we get too excited – simply asking the question is only the first step. We are conscious that there are challenges involved in ensuring that accurate data on sexual identity and gender identity can be collected through a census. The fact that census responses are often completed by one member of a household poses a real barrier to disclosure for those who are not out about their sexual orientation or gender identity to their families.41 Some of the concerns about the limitations of a household form in collecting this information have been addressed by the ONS. Respondents will be able to request a separate access code to complete the census form privately, either on paper or online, without the rest of the household being able to see their responses. 42 However, the ONS also acknowledges in its December 2018 research update that there will likely be under-reporting of the lesbian, gay and bisexual population

37 The Census (England) Regulations 2020 (SI 2020/560), Schedule 2, part 4; The Census (Wales) Regulations 2020, (SI 2020/55 W. 128), Schedule 2, part 4. 38 ONS, Sexual orientation: UK, 2018, 4 October 2017 39 ONS, 2021 Census topic research update: December 2018 40 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018), pp. 39-40 41 Stonewall, Census could become LGBT inclusive (24 May 2016) 42 ONS, Statistical design for Census 2021, England and Wales (October 2020) 21 Commons Library Briefing, 22 October 2020

amongst people aged 16-18 years because the householder will be likely to complete the form on behalf of people in this age group. 43 Past service in the UK Armed Forces There will be a new question identifying veterans of the UK armed forces. The question will ask all respondents aged 16 or over if they have ever served in the UK armed forces (including regulars, reservists and those who served as part of National Service). The question will be worded as follows: Have you previously served in the UK Armed Forces? Current serving members should only tick “no” ⬜ Yes, previously served in Regular Armed Forces ⬜ Yes, previously served in Reserve Armed Forces ⬜ No 44 National-level estimates of the veteran population were published by the Ministry of Defence (MoD) in January 2019. There were an estimated 1.6 million veterans in Great Britain in 2016, based on data from the Annual Population Survey.45 However, there is no reliable data on veterans for small geographic areas. The Commons Library briefing Support for UK veterans (CBP 7693) explains the currently-available data on veterans in more detail. The White Paper reports that there was a user need from central government, local authorities and charities for better data to monitor their performance under the Armed Forces Covenant. The Covenant sets out an agreement between the nation, the government and the armed forces to ensure that members of the armed forces community are supported in their local areas. Local authorities are encouraged to provide assistance with healthcare, education and housing, amongst other things. 46 A research update published by the ONS in October 2017 provides more detail on the development of the question. The ONS tested an alternative approach of linking data from the Veterans Leavers Database to census responses, but found that the database doesn’t adequately cover the population of service leavers for this purpose. Results from this approach are available in the MoD publication Census 2011: Working age UK armed forces veterans residing in England and Wales. Testing found that a question on past service is generally well- understood and considered acceptable to ask. 47

43 ONS, 2021 Census topic research update: December 2018 44 The Census (England) Regulations 2020 (SI 2020/560), Schedule 2, part 4; The Census (Wales) Regulations 2020, (SI 2020/55 W. 128), Schedule 2, part 4. 45 Veterans in this estimate are defined as those who have left the UK Armed Forces and were previously Regular and/or Reserve. 46 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018), pp. 34-35 47 ONS, Update on meeting information needs on the armed forces community (veterans), October 2017 22 Preparing for the 2021 census (England and Wales)

The ONS published a summary of question development for this question in June 2020.

3.3 Repeating questions This section looks at questions that were asked in 2011 and in previous years, but which the ONS plans to amend. For a full list of topics covered in 2011, see page 16. Ethnic group The 2021 census will ask about ethnicity in a similar way to the 2011 census. It will ask the question ‘What is your ethnic group?’, with response tick-boxes grouped under the headings ‘White’, ‘Mixed / multiple ethnic groups’, ‘Asian / Asian British’, ‘Black / African / Caribbean / Black British’ and ‘Other ethnic group’. In Wales, the headings will be slightly different, with ‘Asian Welsh’ and ‘Black Welsh’ included as part of the relevant headings. As well as providing several response tick-boxes under each of these headings, there will also be a write-in box (as there was in 2011). For example, there are four tick-boxes under ‘Asian / Asian British’ as well as the option to write in any other Asian background. The online census will include a ‘search-as-you-type’ function for write-in options that will suggest ethnic groups to users as they type into a box (see section 4.1). 48 The 2021 census will include a new tick-box for the Roma population. The ONS also made a detailed assessment of the need for tick-boxes for the Sikh, Somali and Jewish populations but decided not to recommend that these be included. The ONS published a detailed summary of its process for determining tick-boxes for the ethnic group question in June 2020. New tick-box for Roma respondents The 2021 census will include an additional ‘Roma’ tick-box within the White category in the ethnic group question. It will be listed after the ‘Gypsy or Irish Traveller’ tick-box, which appeared in the 2011 census as well. 49 The ONS cites evidence that there is specific user need for information on the Roma population when developing local lettings policies and planning services, including school places and housing. ONS research found that use of a Roma tick-box, and placing it within the ‘White’ category, was acceptable to Roma focus group participants. The research is described in detail in the ONS’ 2018 research update and in reports published by the contracted research company, Kantar Public. 50

48 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018) 49 The Census (England) Regulations 2020 (SI 2020/560), Schedule 2, part 4; The Census (Wales) Regulations 2020, (SI 2020/55 W. 128), Schedule 2, part 4. 50 Kantar Public, Gypsy Irish Traveller question testing: report findings (August 2018); Kantar Public, Development of the census 2021 ethnic group question: focus groups to consider the addition of possible new tick boxes (August 2018) 23 Commons Library Briefing, 22 October 2020

Current estimates of the size of the UK Roma population come with some uncertainty. Iain Bell, the Deputy National Statistician for Population and Public Policy, told the Women and Equalities Select Committee: Last time round, without the tick box and without any of the engagement with the Roma community, only [730] people wrote in “Roma” under that option. The best estimates we have for the population is they are well in excess of 100,000 51 A 2013 estimate extrapolated from a survey of local authorities put the UK Roma population at nearly 200,000 individuals, with around 194,000 of these living in England and Wales (primarily England). 52 The ONS has said that engagement work with the Roma community will be needed to ensure an accurate result from the census – see section 4.5 for more on this. No tick-box for Sikh respondents The ONS considered the option of including ‘Sikh’ as a tick-box response in the ethnic group question but in the White Paper outlines its decision not to recommend this. There will continue to be a Sikh tick- box in the religion question. A key argument made by those in favour of including a Sikh ethnic group tick-box is that many public bodies monitor ethnicity using the categories set out in the census, which means that without a census tick-box the Sikh community is overlooked. This argument was made by the Sikh Federation UK and Sikh Network in their response to the ONS’ consultation on census topics, 53 and has more recently been made by Preet Kaur Gill MP in a letter to the Guardian: It is a legal fact that the House of Lords ruled in the Mandla v Dowell-Lee case of 1983 that Sikhs are an ethnic group and not simply a religion. As there is not a separate Sikh ethnic tick box in the census, the majority of schools, hospitals, local authorities and other public bodies ignore Sikhs when considering jobs and service provision. It also explains why earlier this month the PM’s race disparity audit totally ignored Sikhs. […] The Office for National Statistics (ONS) cannot ignore the 83,362 Sikhs who in the last census rejected the existing ethnic group categories and ticked “other” and wrote Sikh, or the stakeholders working in the education, health, local government and business sectors that recognise the need for information on Sikhs to plan and make decisions on service provision. 54 The ONS identified other arguments in favour of a Sikh tick-box in its write-up of a 2017 meeting between the ONS and representatives of

51 Women and Equalities Committee, Oral Evidence: Preparations for the 2021 Census, HC 1909, 6 February 2019 52 Brown, P., Scullion, S. and Martin, P. Migrant Roma in the United Kingdom: population size and experiences of local authorities and partners, 2013. 53 Sikh Federation UK and Sikh Network. Response to consultation: The 2021 Census – Initial view on content for England and Wales, 2015. Accessed 15 February 2019 via ONS, 2021 Census topic consultation landing page 54 Gill, P. K. A Sikh ethnic tick box is needed in the UK’s 2021 census. The Guardian, 23 October 2017 24 Preparing for the 2021 census (England and Wales)

the Sikh community. These included concerns that, because the religion question on the census is voluntary, the size of the Sikh community may be underestimated if this measure is used alone. 55 The ONS has carried out and commissioned research into the acceptability of a Sikh tick-box, as well as assessing whether the religion and ethnic group questions are likely to be answered by distinct populations. This research is summarised in a 2017 research report a, 2018 research update, and in the ONS’ detailed summary of its process for determining tick-boxes for the ethnic group question. Focus groups with Sikh participants found that support for an ethnic group tick-box was not universal. The ONS reports that some “younger, second-generation” participants in particular did not find the idea acceptable and did not wish to identify exclusively as ethnically Sikh. Other participants who did wish to express a Sikh ethnic identity did not feel burdened by using a write-in option. 56 A research report summarising the focus groups has been published by the contracted researcher, Kantar Public [PDF]. 57 The White Paper also summarises the findings of quantitative research which suggests that there is very close overlap between respondents who identify as religiously and ethnically Sikh on the census: Additional, quantitative survey findings show there is no evidence that the religious affiliation and ethnic group questions are capturing different Sikh populations. All respondents who stated they were ethnically Sikh (in question versions with or without a specific Sikh response option) also stated their religious affiliation was Sikh. This is in line with findings from the 2011 Census data (where only 1.6% of those who had recorded themselves as ethnically Sikh had a religious affiliation other than Sikh). 58 The ONS has said that it will meet the need for better data about the Sikh community through the census and other data sources: The 2021 Census will continue to include a religion question, with a specific Sikh response option. Flexible data outputs will allow analysis of those who define their religious affiliation as Sikh (through the religion response option) and those who define their ethnic group as Sikh through the use of the “search-as-you-type” capability on the online ethnic group question. ONS will estimate the Sikh population using alternative data sources to assess the numbers who may declare themselves of Sikh background but not through the religion question. ONS will strengthen the harmonisation guidance on the collection of religion alongside ethnicity data across government. ONS will also increase the analytical offering and outputs for all ethnic groups, through flexible outputs. 59

55 ONS, Summary of the meeting with the Sikh community: 23 October 2017 56 ONS, 2021 Census topic research update: December 2018 57 Kantar Public, Development of the census 2021 ethnic group question: focus groups to consider the addition of possible new tick boxes (August 2018) 58 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018), p. 52 59 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018), p. 51 25 Commons Library Briefing, 22 October 2020

The Women and Equalities Committee raised this issue in its oral evidence session on the census in February 2019. The Deputy National Statistician for Population and Public Policy responded: […] what is going on in public service monitoring is that our guidance is clear that we encourage people to look at nationality, religion and ethnicity in the round. However, many public bodies are only picking up the ethnicity. Therefore, simply moving a categorisation from a religion question into it does not tackle the wider problem, which is that actually we should have monitoring of all religions across all public services. The Digital Economy Act helps that. We will be strengthening our guidance in two ways. One is really reinforcing the need to monitor religion. Secondly, we will make it very clear that, for those who do not take forward the religion question […] public sector bodies will have to undertake the due diligence themselves, to make sure they are fully compliant with their duties under that law. 60 In December 2019, the Sikh Federation UK sought judicial review of the planned Census Order (the Census Order is secondary legislation setting out the contents of the census – see section 2.1). The claim was dismissed “on the ground that it is premature, and in breach of parliamentary privilege and the constitutional convention of the separation of powers” – that is, that it came before the Census Order had been laid, and was not seen to justify breaking with the general principle of separation between Parliament and the judiciary. 61 The Sikh Federation UK has since been granted permission for another judicial review hearing, which had not taken place at time of writing. 62 No tick-box for Somali or Jewish respondents The ONS also considered including ‘Somali’ and ‘Jewish’ ethnic group tick-boxes but has decided against recommending these. In the case of the Somali tick-box, the ONS found evidence that there is user need for data on people of Somali origin for service planning. However, ONS research found that some members of the Somali population and some Black Africans felt that the tick-box would single out a single African ethnicity and was therefore unacceptable. The ONS has said that it will look at meeting the need for data through the write- in option and other methods. 63 Research into the acceptability of a Jewish ethnic group tick-box came to a similar conclusion. The ONS found a “strong consensus” amongst focus group participants that a Jewish ethnic group tick-box would not be acceptable. The ONS plans to continue to include ‘Jewish’ as a tick- box response to the religion question. 64

60 Women and Equalities Committee, Oral Evidence: Preparations for the 2021 Census, HC 1909, 6 February 2019 61 R (Gill) v UK Statistics Authority [2019] EWHC 3407 62 Leigh Day, Cabinet Office facing legal challenge over Census 2021 (July 2020) 63 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018), p. 50 64 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018), p. 52 26 Preparing for the 2021 census (England and Wales)

A research report on the focus groups with Somali and Jewish respondents has been published by the contracted researcher, Kantar Public [PDF]. 65 In both cases, the White Paper notes that the open response box will be available for respondents wishing to record their ethnicity as Somali or Jewish. Marital and civil partnership status The 2011 census asked respondents about their marital status, including options for same-sex civil partnerships. Same-sex marriage was legalised in England and Wales in 2013, and opposite-sex couples have been able to register a civil partnership in England and Wales since December 2019. These changes necessitate additional response categories for the 2021 census. The 2021 census will ask respondents what their marital or registered civil partnership status is. In a separate question, it will ask whether the marriage or civil partnership is (or was) to someone of the opposite sex or the same sex. 66 National identity The 2021 census will ask respondents how they would describe their national identity, providing tick-box options for English, Welsh, Scottish, Northern Irish and British as well as a write-in box. 67 This is the same question as in 2011. The ONS has considered whether to include Cornish as a national identity tick-box but has decided not to recommend doing so. It will be possible for respondents to use the write-in space to give their national identity as Cornish, aided by the ‘search-as-you-type’ function (see section 4.1). The main reason for this decision given by the ONS is that user need for data about Cornish people is primarily localised within Cornwall.68 The White Paper states that in the 2011 census, around 83,000 residents wrote in Cornish as their national identity; of these, around 73,000 were living in Cornwall and 10,000 were living elsewhere in England and Wales. The White Paper also acknowledges that Cornish was recognised as a National Minority under the European Framework Convention of the Protection of National Minorities in 2014. 69 Unpaid care The 2021 census will ask respondents whether they provide unpaid care to family, friends, neighbours or others and the amount of unpaid care provided. This question will be amended from the 2011 version in order

65 Kantar Public, Development of the census 2021 ethnic group question: focus groups to consider the addition of possible new tick boxes (August 2018) 66 The Census (England) Regulations 2020 (SI 2020/560), Schedule 2, part 4; The Census (Wales) Regulations 2020, (SI 2020/55 W. 128), Schedule 2, part 4. 67 The Census (England) Regulations 2020 (SI 2020/560), Schedule 2, part 4; The Census (Wales) Regulations 2020, (SI 2020/55 W. 128), Schedule 2, part 4. 68 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018), p. 53 69 HM Government, Press release: Cornish granted minority status within the UK, 2014 27 Commons Library Briefing, 22 October 2020

to make it easier to answer. The ONS has also said that it will amend the response options in order to better align the data with eligibility criteria for Carer’s Allowance. 70 Long-term health and disability The 2011 census asked respondents whether they considered themselves to be ‘long-term sick or disabled’. The White Paper stated that the ONS intends to ask about long-term health and disability in 2021, but with a different question format. 71 The questions used in the autumn 2019 census rehearsal were based on the Government Statistical Service harmonised standard question on disability. 72 Qualifications The 2011 census asked respondents what qualifications they had. According to the White Paper, this information is widely-used but the question has historically been difficult for respondents to answer. The ONS said that it would carry out research to develop a question that is easier to answer. 73 The 2021 census question collects similar information but is phrased differently. 74 Economic activity The 2011 census asked respondents a set of questions designed to determine their economic activity – i.e. whether someone is in employment, or falls into another category such as retired, studying or looking for work. One of these questions in 2011 asked respondents the year in which they last worked, primarily as a means of measuring long-term unemployment. This question was found to be difficult for respondents to answer, and will be replaced with a question about whether or not the respondent has worked in the last 12 months. 75

3.4 Discontinued questions Number of rooms The 2011 census asked households about the number of rooms and the number of bedrooms in their residence. The 2021 census will not include a question about the number of rooms, although it will still ask about number of bedrooms. The ONS instead plans to use administrative data from the Valuation Office Agency (VOA) to estimate the number of rooms at each address (see section 5.1 of this briefing).

70 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018), p. 54 71 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018), p. 55 72 ONS, Census rehearsal household questionnaire – England (H1), [PDF, accessed 5 March 2020] 73 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018), p. 55 74 The Census (England) Regulations 2020 (SI 2020/560), Schedule 2, part 4; The Census (Wales) Regulations 2020, (SI 2020/55 W. 128), Schedule 2, part 4. 75 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018), p. 55 28 Preparing for the 2021 census (England and Wales)

4. How the 2021 census will work

This section sets out plans for 2021 census operations, including preparation and follow up. Further detail is available in the ONS’ Statistical design for Census 2021, England and Wales publication and in its operational planning response to Covid-19.

4.1 An online census The 2021 census will be the first census conducted primarily online in England and Wales. Its target is for 75% of all household responses to be completed online. 76 The White Paper states that a primarily online census will be more cost- effective than a census based on paper forms, partly because online responses are expected to improve the quality of data collected as well as the time taken to process it. The ONS expects online response to be easier for many people, as online services (both government and commercial) have become commonplace. The ONS also plans to enhance the write-in options on census questions by including a ‘search-as-you-type’ facility for the online version of these questions. This will be designed to provide suggestions as users start to type into the write-in boxes, although users will still be able to enter text that doesn’t match any of the suggestions. Paper-based forms will also be available for those who request them – section 4.4 has more detail on this.

4.2 Population covered by the census The ONS plans to collect data on everyone living in England and Wales who is usually resident in the UK, as well as everyone who is present in England and Wales on the date of the census. 77 Household census returns will need to include each person usually resident in the household, as well as anyone who happens to be away on the census date. The census will also ask householders to record basic information about visitors present at their address, including the visitors’ usual address. 78 Residents who live in communal establishments rather than households will be counted as well – this includes places such as hospitals, residential care homes, and hotels. The White Paper states that: Special arrangements will be made to enumerate residents in prison service establishments and immigration detention centres, on ships in ports and in military camps, and other population

76 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018), p. 78 77 Ibid., pp. 26-27 78 Ibid., pp. 26-27 29 Commons Library Briefing, 22 October 2020

groups not resident in private households, such as persons travelling with fairs. 79

4.3 Covid-19 and the census The Covid-19 pandemic poses operational challenges to census-taking, with the potential to affect both the preparation phase (e.g. engaging with community groups), and the census itself (e.g. carrying out face-to- face work to support people to complete returns). In July 2020, the Scottish Government announced that it would postpone Scotland’s census for a year because the Covid-19 outbreak had affected much of its preparation activity. The Registrar General and Chief Executive of the National Records of Scotland (NRS) commented: The census is a unique count of every household in Scotland and in the next census, for the first time, the majority of people will be encouraged to complete their questionnaire online. This will require a significant field force and face-to-face engagement in an effort to support digital participation, engage with the public and raise awareness of census across the population. The census is a major logistical operation and the twelve months running up to a census are vital in planning and testing the effectiveness and security of systems and processes. COVID-19 restrictions have delayed or prevented many of these key activities, including engagement with key organisations, from taking place over recent months.80 The ONS still plans to carry out the census in March 2021 in England and Wales, and NISRA plans to do the same in Northern Ireland. In October 2020 the ONS published an operational planning response that sets out how it plans to address the impact of the pandemic on census operations. The document sets out what it identifies as the main potential impacts, which include: • Social distancing restrictions could affect the ability of field staff to work face-to-face in community engagement activities or with households. • Social distancing and travel restrictions could affect supply and distribution of materials, and the ability of central and field staff to work. • Respondents may be less likely to answer their door to field staff, take part in in-person community engagement, etc. • The economic impact of the pandemic could affect field staff recruitment and/or the ability of contracted suppliers to complete work. 81

79 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018), pp. 26-27 80 News release – Scotland’s Census to be moved to March 2022, 17 July 2020 [Accessed 7 October 2020] 81 ONS, Operational planning response to the coronavirus (COVID-19) for Census 2021, England and Wales, 1 October 2020 30 Preparing for the 2021 census (England and Wales)

The document explains how the ONS has changed its plans and made contingency plans, in order to ensure that the census can still be delivered. Some of these steps are noted in the sections below. The pandemic could also affect how respondents answer the census. The ONS acknowledges this in its statistical design document: The coronavirus (COVID-19) has changed how people are living, working and studying. We are working to understand how respondents interpret the questions, for example, on location of people (particularly university students), health and caring responsibilities, workplace and travel to work, and how this could impact the outputs from the census through engagement with stakeholders and topic experts. While the exact questions and response options have been finalised, we are considering whether additional guidance may be needed to help respondents interpret the questions in different scenarios, for example, on students’ term-time address. 82

4.4 Operation of the census Staffing Field staff will be employed to support the operation of the census: the White Paper says that this “will consist of full- and part-time staff with at least 17,000 census field officers being recruited”. Field officers’ duties will include making contact with non-responding households and assisting respondents with online and paper completions. Census engagement managers and liaison staff will also be recruited (see section 4.5 for more on these roles). The ONS plans to recruit enough Welsh-speaking field staff in Wales that Welsh speakers are supported in completing the census. Recruitment, training and payment will be outsourced to a contractor, but the field staff will be ONS employees. 83 Identifying and contacting respondents The ONS will use an address register (based on Ordnance Survey’s Address Base Premium product and supplemented with administrative data) to identify and contact every household and communal establishment for the census. Initial contact will be made by post around three weeks before the census date, letting households know that the census is coming.84 The ONS plans to reach households by posting them an invitation to complete the census form online, along with a Unique Access Code (UAC). This will be delivered at least a week before census day. 85 The ONS acknowledges that some respondents will need to use a paper questionnaire to answer the census. Paper forms will be sent to anyone who requests them. Additionally, in some geographic areas where the

82 ONS, Statistical design for Census 2021, England and Wales, 1 October 2020 83 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018), pp. 76-77 84 ONS, Statistical design for Census 2021, England and Wales, 1 October 2020 85 Ibid. 31 Commons Library Briefing, 22 October 2020

ONS judges that paper forms will be more successful, residents will initially receive paper forms rather than invitations to respond online. Respondents in those areas will also be given a UAC and so will still be able to answer online. 86 Anyone will be able to request a separate UAC or paper form if they wish to complete the census privately, without the rest of the household being able to see their responses. Individual responses will take precedence over answers provided for the same person on a household form. 87 Additional services will be offered to support people to respond to the census online. The ONS plans to set up Census 2021 Support Centres in locations such as libraries and community centres. Covid-19 safety measures will be put in place at in-person Support Centres, and the ONS will develop contingency options in case centres have to close.88 Online and telephone-based help will also be available via a Contact Centre, and respondents will be able to complete their census return by telephone if they need to. 89 The method of contact and completion for communal establishments will vary depending on the type of establishment. Field staff will liaise with local authorities and communal establishments to establish the best approach in each case. 90 Coverage and non-response The ONS aims to achieve a 94% response rate across England and Wales, and a response rate of at least 80% in each local authority. Further targets are set for smaller areas and are explained in the ONS’ statistical design document. 91 The ONS will track response rates in real time in order to allocate resources for follow-up contact. Households that do not respond initially might receive a paper questionnaire as a reminder, or a visit or phone call from a member of field staff. 92 Field staff visits will be carried out in line with social distancing guidance, and may be replaced with phone calls if necessary. 93 There is a statutory requirement to complete a census return. The exact definition of the population who are subject to this requirement is set out in the Census (England and Wales) Order 2020. According to the White Paper: Persons refusing to comply with the statutory requirement will be offered encouragement and assistance to do so, but if they continue to refuse, may be liable to legal proceedings. This may

86 ONS, Statistical design for Census 2021, England and Wales, 1 October 2020 87 Ibid. 88 ONS, Operational planning response to the coronavirus (COVID-19) for Census 2021, England and Wales, 1 October 2020 89 ONS, Statistical design for Census 2021, England and Wales, 1 October 2020 90 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018), pp. 78-80 91 ONS, Statistical design for Census 2021, England and Wales, 1 October 2020 92 Ibid. 93 ONS, Operational planning response to the coronavirus (COVID-19) for Census 2021, England and Wales, 1 October 2020 32 Preparing for the 2021 census (England and Wales)

include prosecution, possibly resulting in a fine (the maximum level of which is currently £1,000) and a criminal record. 94 Prosecution for non-completion of a census form has been relatively rare. There were 286 cases taken to court for non-completion following the 2011 census, of which 270 resulted in convictions. 95 The ONS will estimate levels of under-counting and over-counting in different groups through the Census Coverage Survey (CCS). The CCS is carried out by taking a random sample of postcodes, and sending interviewers to every household in the postcode (a total sample of around 350,000 households).96 Census rehearsal A census rehearsal took place on 13 October 2019, with the aim of ensuring that the procedures put in place for collecting census responses work at scale. The rehearsal took place in four local authority areas: Carlisle, Ceredigion, Hackney and Tower Hamlets. These local authorities were chosen to include “a cross-section of the population and types of housing” in testing. 97 All households and communal establishments in these rehearsal areas were asked to complete census returns. The ONS published an evaluation report on the 2019 census rehearsal on 1 October 2020. The rehearsal was generally considered to be successful, particularly in terms of online responses.

4.5 Engagement and support The ONS is planning to carry out engagement and support work in order to help meet its completion targets, and ensure that groups who are at risk of being under-counted are able to complete the census. The White Paper describes plans for a publicity campaign aimed at informing householders about when and how to complete the census. It says that the messaging “will explain the purpose and value of the census, and give reassurance about confidentiality and data security. This will include messages about the legal requirement to complete the census”.98 The ONS also plans to provide support services to help people fill in the census, namely a Contact Centre that can be reached online or by phone and in-person Census 2021 Support Centres operating at some locations, offering in-person assistance with filling in the census online. In Wales, the questionnaire and supporting materials will be available in Welsh and English. Translation leaflets will be available in over 40

94 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018), p. 82 95 ONS, Freedom of Information response: Non-compliance procedures relating to the 2011 census (December 2015) 96 ONS, Statistical design for Census 2021, England and Wales, 1 October 2020 97 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018) p. 24 98 Ibid., p.22 33 Commons Library Briefing, 22 October 2020

languages, and interpretation services will be provided via the Contact Centre. 99 The ONS has also said that it is considering how best to provide support in British Sign Language, and is testing the online completion process to ensure accessibility. 100 Guidance leaflets about the census will be available in Braille, large print, British Sign Language and in easy-read versions. 101 The ONS plans to liaise with local authorities and community groups in order to better reach groups that have been under-counted in the past. There are plans to employ field staff who are representative of the communities that the ONS seek to reach, as well as working with local and national community groups. 102 There has been particular interest in the ability of the ONS to adequately reach the Roma, Gypsy and Irish Traveller communities. The issue was raised by the Women and Equalities Select Committee during its oral evidence session on preparations for the 2021 census. Eddie Hughes MP: On the issue of the Roma ethnicity classification, how do you intend to ensure that people are aware of which tick box to use, and that the headcount is accurate? Iain Bell: First, we know that we have already begun engagement with these groups. In the research and stakeholder engagement in order to determine the tick boxes, we have already formed links with national organisations representing the Roma community and with local organisations. We have already made links in Sheffield and elsewhere. The approach is under development, but it will be very similar to what we did in 2011 with Gypsy or Irish Travellers. We know local authorities are good for helping to tell us where the communities are, but both we and local authorities do suffer a little, in that we have association with Government and not all these groups fully trust Government. The approach we use is to employ community outreach workers and community outreach census teams, who then go in and work with the communities in order to get the completion in. […] Chair: It is fair to say that the last census, in terms of its ability to be able to estimate the numbers of Gypsies and Travellers, was felt to be a vast underestimation. Are you really right to be using similar tactics this time round? How are you going to make sure it is not just another huge underestimation of numbers in the Roma community this time? Iain Bell: It is fair to say we know that the undercount in Roma was even more severe. […] We are always happy to work with communities to ensure we can get this form of engagement further still. 103

99 ONS, Statistical design for Census 2021, England and Wales, 1 October 2020 100 ONS, Equality impact assessment for the 2021 Census (March 2020), pp. 8-9 101 ONS, Census website, Accessibility [Accessed 11 March 2020] 102 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018) p. 82-84 103 Women and Equalities Committee, Oral Evidence: Preparations for the 2021 Census, HC 1909, 6 February 2019 34 Preparing for the 2021 census (England and Wales)

Concerns have also been raised in Parliament about the ability of the census to capture data about homeless people. Baroness Hayter of Kentish Town raised this issue at the second reading of the Census (Return Particulars and Removal of Penalties) Bill: For the census to be complete […] it needs to count those who are homeless as well as those who are housed. This is important both to measure the impact of demographic or policy changes and to plan services for this vulnerable group […] but, as we know, this group is currently underrepresented on the returns. Last year Shelter met the relevant ONS team and made suggestions to improve the situation. In particular, it argued that the key to achieving an accurate count by including homeless people is really just trying harder to reach them. That might mean providing extra reassurance to the homeless about dealing with officials and stressing to every local authority the importance of that. Without accurate numbers, there is little chance that services to help those experiencing homelessness will be fairly and adequately delivered. 104 The ONS has said that it will work with homelessness organisations to ensure that as many homeless people as possible are recorded on the census: All people who are homeless will be able to complete the census. For some, this will be through a form at an address where they are temporarily staying, for example, with a friend or in a hostel. Others will be able to use the address of an establishment that they are in contact with (for example, a day or night shelter). Early in 2020, ONS will host a session to rehearse the enumeration of people who are homeless in England and Wales using a range of scenarios. This will involve stakeholders who have practical experience of engaging with homeless people and working with other places homeless people come into contact with (for example, food banks and soup kitchens). Following this work, ONS will release more information about its approach to maximising response. 105 The ONS acknowledges that the Covid-19 pandemic poses a risk to community engagement continuing as planned, but aims to switch to an increased focus on online engagement: Where our plans had envisaged community engagement to principally be carried out face-to-face and in person, the risks of infection, restrictions and public perception mean that there could be significantly less scope for this form of engagement. However, we have found that many community groups with whom we planned to engage have moved online. Design changes have already been implemented to enable an increased focus on online engagement with these communities as a mitigation if required, should face-to-face engagement opportunities be lost because of coronavirus restriction […] To facilitate this online engagement, we are equipping our community engagement teams with the ability to engage with communities securely through social media.106

104 HL Deb 13 May 2019 c1431 105 ONS, Improving homelessness and rough sleeping statistics across the UK (17 September 2019) 106 ONS, Operational planning response to the coronavirus (COVID-19) for Census 2021, England and Wales, 1 October 2020 35 Commons Library Briefing, 22 October 2020

4.6 Costs The White Paper gives the estimated costs of the wider Census and Data Collection Transformation Programme. The programme has three aims: • a predominantly online census of all households and communal establishments with special care taken to support those who are unable to complete the census online • evidence to enable a decision about the future provision of population statistics after 2021 • modernise ONS’s data collection activities to provide improved and expanded population statistics, which will deliver efficiencies through the greater use of administrative data and survey data collected online instead of interview and paper based data collection. 107 The White Paper describes the cost of the programme as follows: Provision of around £900 million for the total cost of the Census and Data Collection Transformation Programme, and Digital Transformation Programme has been made for the period 2015- 2026. Provision for later years will be subject to future spending reviews and will cover the final stages of data and output processing, dissemination and a number of subsequent output services and releases. The programme is subject to annual business case refreshes which are approved by HM Treasury and the Cabinet Office. 108 The census itself – excluding the costs of the wider transformation programme – will cost around £800 million, according to oral evidence given by the Deputy National Statistician, Iain Bell, to the Women and Equalities Select Committee. 109

107 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018), p. 124 108 PQ 207229 [on Travellers: Census], 10 January 2019 109 Women and Equalities Committee, Oral Evidence: Preparations for the 2021 Census, HC 1909, 6 February 2019 36 Preparing for the 2021 census (England and Wales)

5. Data processing and outputs 5.1 Data processing, analysis and security Processing census data The ONS will use a response management system to keep track of which addresses have returned census forms, and which need follow- up. Once completed census forms are received by ONS, the data will be processed and adjusted to account for errors and non-response. The ONS recognises that the census will not achieve a 100% response rate – it is aiming for a 94% response rate across England and Wales, and a response rate of at least 80% in each local authority. This means that the data collected in the census will need to be adjusted to make sure that it accurately reflects the population. A separate Census Coverage Survey (CCS) will independently contact a sample of addresses in order to estimate how many people have not completed the census and what their characteristics are (see section 4.4). 110 Data security The Statistics and Registration Service Act 2007 requires UKSA and the ONS to protect the confidentiality of data collected during the census. Access to census records will be restricted to employees and contractors working for UKSA (and by extension, the ONS) on the census. Employees and contractors will be security-checked and will be liable for prosecution if they unlawfully disclose personal census information. Researchers who have been accredited by UKSA will also have access to anonymised census records under secure conditions. Government policy to date has been to keep census records closed to the public for 100 years, and the intention is for this to continue – this means that records from the 2021 census will not be available to the public until January 2122. Linking to administrative sources The ONS has committed to using administrative data (i.e., data already held by the government for other purposes) to enhance census data. This approach was proposed by the National Statistician and endorsed by the government in 2014 (see section 1.4). The ONS plans to link census data with administrative data from the Valuation Office Agency (VOA) to estimate the number of rooms at an address. The ONS has carried out feasibility research into this approach using 2011 Census and VOA data. In July 2020, it published a summary of quality assurance undertaken on the VOA data in which it concludes that the data is suitable for use for this purpose. In October 2020, it

110 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018) pp. 97-98 37 Commons Library Briefing, 22 October 2020

published a further report on how it intends to address missing values, and summarised its approach in a blog post. 111,112 An income question was considered for the 2021 census form, but the ONS decided not to include it as testing found that people were less likely to respond to a census that asks about income. The ONS was also concerned about participants’ ability to accurately report their income. The ONS instead plans to work with tax and benefits data (held by HM Revenue and Customs and the Department for Work and Pensions, respectively) to develop “census-style income data” for small areas. 113 Initial experimental statistics based on this approach were published in December 2018.

5.2 Dissemination The ONS aims to publish some headline national and local authority level outputs within one year of the census date, and all other “main data” within two years. 114 The ONS will publish headline data and reports on specific topics. In addition, it plans to develop an online, flexible dissemination system that will let users select the geography, population base and variables that they require (while maintaining confidentiality). This is in contrast to outputs from previous censuses, which have been limited to specific pre-constructed tables. 115 When statistics are made available for specific populations in small geographic areas, there is a risk that individuals or households will be identifiable from their data. The ONS proposes to mitigate this risk through a number of approaches, including by systematically introducing random noise into the available data (to an extent that doesn’t significantly affect data quality) and limiting some data categories to aggregates (e.g. age bands). 116 More information about the ONS’ progress on dissemination is available on the ‘Census 2021 products’ pages of the ONS census website.

111 ONS, Estimating the number of rooms in Census 2021: an update on imputation methods for Valuation Office Agency data, 31 July 2020. 112 ONS, Census 2021 – For the first time the ONS is using administrative data to count number of rooms, 31 July 2020 113 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018) p. 64 114 ONS, Statistical design for Census 2021, England and Wales, 1 October 2020 115 ONS, Statistical design for Census 2021, England and Wales, 1 October 2020 116 HM Government, Help shape our future: the 2021 census of population and housing in England and Wales, (December 2018) pp.106-107 38 Preparing for the 2021 census (England and Wales)

Appendix 1: History of census topics

The table below summarises topics covered by censuses between 1841 and 2021. It is intended as a broad summary rather than a detailed explanation of the questions asked at each census – see p. 15 for a full list of 2021 census topics.

Census topics in England and Wales, 1841-2011

1841 1851 1861 1871 1881 1891 1901 1911 1921 1931 1951 1961 1966 1971 1981 1991 2001 2011 2021 Age Sex Gender identity Marital / civil partnership status Fertility within marriage Marriage duration

Birthplace National identity Main language Passport Migration Ethnicity Religion Sexual orientation

Occupation Economic position Journey to work Industry Time since last employment Size of workforce Place of work Supervisor status Qualifications Veteran status

Infirmity Self-assessed general health Limiting long-term illness Provision of unpaid personal care

Usual residence Second address Tenure Number of rooms Lowest floor level Central heating / type of heating Access to motor vehicles

Notes: An additional census took place in 1966, based on a sample of the population. Sources: Office for National Statistics (1993), The 1991 Census: A User’s Guide As cited in House of Commons Library (2013), 2011 Census of Population: background HM Government (2018), Help shape our future: the 2021 census of population and housing in England and Wales

About the Library The House of Commons Library research service provides MPs and their staff with the impartial briefing and evidence base they need to do their work in scrutinising Government, proposing legislation, and supporting constituents. As well as providing MPs with a confidential service we publish open briefing papers, which are available on the Parliament website. Every effort is made to ensure that the information contained in these publicly available research briefings is correct at the time of publication. Readers should be aware however that briefings are not necessarily updated or otherwise amended to reflect subsequent changes. If you have any comments on our briefings please email [email protected]. Authors are available to discuss the content of this briefing only with Members and their staff. If you have any general questions about the work of the House of Commons you can email [email protected]. Disclaimer This information is provided to Members of Parliament in support of their parliamentary duties. It is a general briefing only and should not be relied on as a substitute for specific advice. The House of Commons or the author(s) shall not be liable for any errors or omissions, or for any loss or damage of any kind arising from its use, and may remove, vary or amend any information at any time without prior notice. The House of Commons accepts no responsibility for any references or links to, BRIEFING PAPER or the content of, information maintained by third parties. This information is Number 8531 provided subject to the conditions of the Open Parliament Licence. 22 October 2020